DATA SOURCES: Two reviewers independently conducted a robust search of MEDLINE PubMed, EMBASE, and the Cochrane Database of Controlled Clinical Trials from their inception to January 2016; bibliographies of identified articles were screened for additional relevant articles.
STUDY SELECTION: Randomized controlled trials, quasi-randomized controlled trials, nonrandomized controlled trials, controlled before-and-after studies, and prospective or retrospective parallel cohort studies reporting frequency of bleeding events were reviewed if they compared one or several antiplatelet drugs with any oral anticoagulant, included patients aged 65 years and older (ie, 70% or more of all included patients were ≥65 years, or the studies reported separate data for this age group), or followed patients for 6 months or more. Studies on post- or perioperative bleeding complications and patients receiving combined treatment with anticoagulant and antiplatelet drugs were excluded. Patients younger than 65 years were excluded.
DATA EXTRACTION AND SYNTHESIS: Two authors independently assessed every study for inclusion, assessed its risk of bias, and extracted data with predefined forms. The main outcome analyzed was major bleeding, defined as clinically overt bleeding involving a critical anatomic site or associated with death, permanent disability, a decrease in hemoglobin concentration greater than 2 g/dL, or a patient’s requiring transfusion of greater than or equal to 2 units of whole blood or packed RBCs. Secondary outcomes were intracranial bleeding, any severity of bleeding, and all-cause mortality. Outcomes were analyzed at 12 months. The risk of bias was assessed with the tool developed by the Cochrane Collaboration. Subgroup analyses were predefined by age (patients 65 to 79 years, or 80 years or older), by indication for anticoagulation (atrial fibrillation or other indications), and by aspirin dose (less than or greater than 100 mg/day).